Complex Task Scope, Steps, and Acceptance Planning

Problem It Solves

Complex multi-step tasks often fail not because a step is impossible, but because the agent loses the original goal, repeats searches, or scatters error context across a long session. AI Maestro Planning is aimed at tasks with 3 or more steps, research projects, and feature work that requires repeated tool calls. It keeps goals, findings, and progress in persistent Markdown files so the working state survives context loss.

How It Works

The skill enforces a plan-before-action loop and maintains three files: task_plan.md for goals, phases, decisions, and errors; findings.md for research results, resources, and key discoveries; and progress.md for session logs, test results, and status. Key rules include:

• Plan first: create task_plan.md before starting complex work.
• Read before decide: re-read the plan before major decisions.
• Update after act: mark phase changes and record what changed.
• Save findings early: after every two search/browse operations, write discoveries to findings.md.
• Log failures: record each error with attempt number and resolution, then change approach instead of repeating the same failure.

Boundaries and Notes

It is useful when state needs to persist across steps, but it is not intended for one-shot questions or trivial tasks that do not require tracking. If an action fails repeatedly, the skill follows a 3-strike escalation: diagnose the root cause, try a different method, then question assumptions and search for similar issues before escalating to the user with all attempts documented.